trailhq/Trail-Code ? reverse-engineered prompt
Reverse engineered prompt
Build me a TypeScript command line tool that helps coding agents understand a codebase better and stop relearning the same repo every session.
I want it to scan a project, build a local code map from the files, and use that map to feed the most relevant context into agents like Claude Code, Cursor, Codex, and Gemini. It should also let me save useful learnings from each session so the next run can reuse them, and it should update the repo’s agent instructions files when needed so the agent remembers team rules instead of forgetting them.
Make the setup feel very simple, something like install once, run an init command, and it wires itself into the repo. I also want commands to rebuild the map, search across the codebase, inspect how the graph was built, and view it somehow if possible. Keep it local by default, work well in monorepos, and use the current docs online if you need to check best practices.
Are you gonna build this?
make sure you review the code using coderabbit